Citation for Published Item: the Protozoan Inositol Phosphorylceramide Synthase: a Novel Drug Target Which Defines a New Class of Sphingolipid Synthase

(2006) 'The protozoan inositol phosphorylceramide synthase : a novel drug target that denes a new class of sphingolipid synthase. (2006) 'The protozoan inositol phosphorylceramide synthase : a novel drug target that denes a new class of sphingolipid synthase.', Journal of biological chemistry., 281 (38). pp. 28200-28209 c the American Society for Biochemistry and Molecular Biology" Additional information: Use policy The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for personal research or study, educational, or not-for-prot purposes provided that: • a full bibliographic reference is made to the original source • a link is made to the metadata record in DRO • the full-text is not changed in any way The full-text must not be sold in any format or medium without the formal permission of the copyright holders. Please consult the full DRO policy for further details. Sphingolipids are ubiquitous and essential components of eukaryotic membranes, particularly the plasma membrane. The biosynthetic pathway for the formation of these lipid species is conserved up to the formation of sphinganine. However, a divergence is apparent in the synthesis of complex sphingolipids. In animal cells, ceramide is a substrate for sphingomyelin (SM) production via the enzyme SM synthase. In contrast, fungi utilise phytoceramide in the synthesis of inositol phosphorylceramide (IPC) catalysed by IPC synthase. Due to the absence of a mammalian equivalent, this essential enzyme represents an attractive target for anti-fungal compounds. In common with the fungi, the kinetoplastid protozoa (and higher plants) synthesize IPC rather than SM. However, orthologues of the gene believed to encode the fungal IPC synthase (AUR1) are not readily identified in the complete genome databases of these species. By utilising bioinformatic and functional genetic approaches, we have isolated a functional orthologue of AUR1 in the kinetoplastids, causative agents of a range of important human diseases. Expression of this gene in a mammalian cell line led to the synthesis of an IPC-like species strongly indicating that IPC synthase activity is reconstituted. Furthermore the gene product can be specifically inhibited by an anti-fungal targeting IPC synthase. We propose that the kinetoplastid AUR1 functional orthologue encodes an enzyme which defines a new class of protozoan sphingolipid synthase. The identification and characterisation of the protozoan IPC synthase, an enzyme with no mammalian equivalent, will raise the possibility of developing anti-protozoal drugs with minimal toxic side-affects.

The evolutionarily divergent insect vector-borne protozoan parasites of the order Kinetoplastidae cause a range of human diseases, including the leishmaniases (a broad spectrum of infections caused by the Leishmania species), Chagas disease (Trypanosoma cruzi) and African sleeping sickness (Trypanosoma brucei). These infections are of increasing prevalence in developing countries where mortality is often affected by poor access to health care. Many of the drugs available to treat these diseases are expensive, difficult to administer, or toxic, and no effective vaccines are available (1,2). These facts make the discovery of new anti-kinetoplastid therapeutic targets and compounds of paramount importance (www.who.int/tdr).
Sphingolipids are amphipathic lipids that have a sphingoside backbone with a long-chain fatty acid and a polar alcohol as attachments. Ceramide is an unmodified sphingolipid that functions as a key player in cell signaling and as a precursor of complex sphingolipids, which are major components of the outer leaflet of eukaryotic plasma membranes thought to be involved, together with sterols, in the formation of microdomains known as lipid rafts (3). These rafts have been proposed to function in a diverse array of processes from the polarized trafficking of lipid-modified proteins (4) to the assembly and activation of signal transduction complexes (5,6). Unlike mammalian cells, which synthesize SM, 3 the primary complex sphingolipid species in Leishmania sp. and Trypanosoma sp. is inositol phosphorylceramide (IPC) (7). Similarly, in fungi and plants, IPC and its glycosylated derivatives are abundant complex sphingolipid species (7,8) (see Fig. 1). IPC is produced by the transfer of a phosphorylinositol head group from phosphatidylinositol to ceramide, a process catalyzed by IPC synthase (9). In the fungi, IPC synthase or a subunit of this enzyme is believed to be encoded by the AUR1 gene that was first identified in Saccharomyces cerevisiae and subsequently in several pathogenic species (10). As in the protozoan parasites, the search for specific drug targets in the pathogenic fungi is hampered by the conservation of biochemical processes between the eukaryotic microbe and its eukaryotic host. Consequently, as for anti-protozoan therapies, many of the available anti-fungals exhibit toxic side effects. IPC synthase has no functional equivalent in mammals. As such, this enzyme activity has been established as a target for the development of anti-fungal compounds (11), a number of which have been characterized previously (12).
No homologue of the fungal AUR1 gene has been found in either the complete protozoan or plant genome sequence data bases. However, given the efficacy of known yeast inhibitors against the plant IPC synthase activity, it appears likely that the enzyme is conserved in plants at least (13). The biosynthesis of sphingoid base and ceramide, precursors of complex sphingolipids, are non-essential for Leishmania pathogenesis in an animal model system (14). However, it is known that intramacrophage amastigotes possess an active IPC synthase presumably utilizing precursors from alternative sources (15). Indeed, it has been shown that Leishmania donovani stimulates host macrophages to up-regulate the production of ceramide, a precursor of IPC and a substrate of IPC synthase ( Fig. 1) (16). Moreover, sphingolipid synthesis is essential for replication of pathogenic bloodstream form T. brucei. 4 IPC synthase activity has been characterized in T. cruzi (17), where it has been suggested that IPC synthesis is important for infectivity (18). In this context, it should be noted that IPC synthase is believed to play a pivotal role in the pathogenesis of the fungus Cryptococcus neoformans, which similar to Leishmania sp. and T. cruzi, resides in an acidic macrophage phagolysosome. Genetically induced down-regulation of the C. neoformans AUR1 confers an in vivo growth defect through a pH-dependent mechanism (19).
Here we have described the identification and characterization of an AUR1 functional orthologue in Leishmania major, the putative protozoan IPC synthase (LmIPCS). It is proposed that this protein defines a new class of eukaryotic sphingolipid synthases. In addition, the development of an ex vivo assay system has allowed the identification of an inhibitor with specific activity against LmIPCS. This work will facilitate the rational development of inhibitors against a protozoan enzyme with no mammalian equivalent, leading to the prospect of anti-protozoal compounds with minimal toxic side effects.

EXPERIMENTAL PROCEDURES
Selection and Cloning of a Candidate AUR1 Orthologue-Candidate genes were identified in a motif search of the genome sequence data bases of L. major, T. brucei, and T. cruzi using the sequence motif H(YFWH)X 2 D(VLI)X 2 (GA)X 3 (GSTA) (Sanger GeneDB). Candidates were then further selected using the additional criteria described under "Results." The Leishmania Candidate Coding Sequence-LmjF35.4990 was amplified from genomic DNA (14) using Pfu polymerase (Stratagene) and the primer pairs 5ЈLmIPCSBamHI (cgc gga tcc ATG ACG AGT CAC GTG ACA GC) and 3ЈLmIPCS*HindIII (ccc aag ctt TTA GTG CTC AGG CAA AGC CGC CG); and 5ЈLmIPCSBamHI and 3ЈLmIPCSHindIII (ccc aag ctt GTG CTC AGG CAA AGC CGC CG). The products were cloned into the yeast and mammalian expression vectors pRS426MET25 and pcDNA3.1/Myc-HisA (Invitrogen) creating pRS426 LmIPCS and pcDNA3.1A LmIPCSMyc-His, respectively. The S. cerevisiae AUR1 gene was similarly amplified from genomic DNA (Invitrogen) using the primer pair AURFXbaI (cat aga tct aga ATG GCA AAC CCT TT TTC GAG ATG G) and 3ЈScIPCS*HindIII (ccc aag ctt TTA AGC CCT CTT TAC ACC TAG TGA CG) and the product cloned into pRS426MET25 to create pRS426 ScAUR1. Cloning sites are shown in lower case with Leishmania and S. cerevisiae sequences in upper case letters.
Construction of YPH499-HIS-GAL-AUR1 Auxotrophic S. cerevisiae Strain-The YPH499-HIS-GAL-AUR1 S. cerevisiae strain was constructed in YPH499 (Mat a; ura3-52; lys2-801amber; ade2-101ochre; trp1-63; his3-200; leu2-1) (Stratagene) by bringing the expression of the yeast AUR1 gene under the control of the stringently regulated GAL1 promoter that is repressed in the presence of glucose (20). The AUR1 promoter in the yeast genome was exchanged by a selection marker/promoter HIS/GAL1 cassette using a previously described methodology (21). The primer sequences for amplification of the HIS/GAL cassette were chosen as follows: (a) sequence for integration upstream of the coding region (nucleotides Ϫ200 to Ϫ150) AuriHISGalS, GGT AGT TGG TTA GTC CGA TCG CTC ACT TTT GGT TGT TGT TAA GTA CTT CAG GGC GAA TTG GAG CTC CAC; (b) sequence for integration at the initiation codon (nucleotides ϩ1 to ϩ50) AuriHISGalAS, CAG TTT GGA GGT CTC TCT GAT AGA AAC CAT CTC GAA AAA GGG TTT GCC ATG GGG ATC CAC TAG TTC TAG. The numbers indicate the nucleotide positions in the S. cerevisiae DNA sequence, with the adenosine of the ATG initiation codon being defined as position ϩ1. The 19-bp sequences at the 3Ј ends of these oligonucleotides, which are homologous to the sequences of the vector pGAL/HIS3 (22) and serve as a template for amplification of the GAL1/HIS3 cassette, are underlined. Transformation into the haploid YPH499 strain, selection on minimal medium lacking histidine but containing galactose, and confirmation of the insertion of the HIS-GAL fragment were performed as previously described (23). YPH499-HIS-GAL-AUR1 was maintained in SGR medium (4% galactose, 2% raffinose, 0.17% Bacto yeast nitrogen base, 0.5% ammonium sulfate) with galactose/raffinose rather than nonpermissive dextrose as the carbohydrate source. For rapid cultivation of the mutant, YPGR medium (1% yeast extract, 2% peptone, 4% galactose, 2% raffinose) was routinely used.
Identification of Leishmania AUR1 Orthologue-The YPH499-HIS-GAL-AUR1 S. cerevisiae strain was transformed with pRS426 ScAUR1 or pRS426 LmIPCS and functionally complemented transformants selected on non-permissive SD medium (0.17% Bacto yeast nitrogen base, 0.5% ammonium sulfate, and 2% dextrose) containing the nutritional supplements necessary to allow selection of transformants.
Sequence Analyses-Based on primary sequence, orthologues of LmIPCS (LmjF35.4990) were found by BLAST search of the T. brucei and T. cruzi data bases (Sanger GeneDB) using WU-BLASTp. Sequence alignments were made using ClustalW (24). Phylogenetic analyses were performed on the edited alignments using maximum parsimony (PHYLIP (Phylogeny Inference Package, version 3.5c). Topology predictions were performed using the PHD package (25).
Expression of LmIPCS in a Mammalian Cell Line-Human ARF1 was amplified from lymph node cDNA (Clontech) using the primers hARF1F (CCT GTC CAC AAG CAT GGG GAA CAT) and hARF1R (CCT TCT GGT TCC GGA GCT GAT TGG). The fragment was cloned into pcDNA3.1/ CT-GFP-TOPO (Invitrogen) to produce the construct pcDNA3.1 hARF1GFP. The cell line HEK293 was grown in Dulbecco's modified Eagle's medium, containing 10% fetal calf serum, 100 units/ml penicillin, and 100 units/ml streptomycin at 37°C in a humidified atmosphere with 5% CO 2 .
For transient transfection, cells were seeded in 6-well plates Ϯ glass coverslips at a cell density of 2 ϫ 10 5 cells/well and allowed to grow for 24 h. The cells were then transfected with pcDNA3.1A LmIPCSMyc-His, pcDNA3.1 hARF1GFP, pcDNA3.1/Myc/His/lacZ, or pcDNA3.1A vector alone using FuGENE 6 transfection reagent (Roche Applied Science). For each well, 3 l of FuGENE reagent was combined with 2 g of plasmid DNA. For co-transfections, 6 l of FuGENE was used per well with 2 g of each construct.
Immunofluorescence-36 h post-transfection HEK293 cells were fixed with 4% paraformaldehyde (w/v) for 45 min at room temperature. Expressed LmIPCS was detected by indirect immunofluorescence. The cells were washed in PBS, permeabilized in 0.5% Triton X-100/PBS (v/v) for 10 min, and then blocked in 10% fetal calf serum/PBS (v/v) for 1 h. The samples were incubated with rabbit polyclonal anti-Myc (Abcam, 1:200 dilution) in blocking solution for 1 h. After washing in PBS, cells were incubated in Alexa Fluor 633 goat anti-rabbit IgG (Invitrogen, 1:250) in blocking solution for 1 h. After washing in PBS, coverslips were mounted on slides using Vectashield with 4Ј,6diamidino-2-phenylindole (Vector Laboratories). Samples were visualized by confocal microscopy using a Zeiss LSM 510 Meta with a Plan-Apochromat 63ϫ/1.4 Oil DIC I objective lens. Images were acquired using LSM 510, version 3.2, software (Zeiss).
Logarithmic phase yeast (unless stated otherwise) was harvested and incubated at 10 A 600 /ml in 0.25 ml of inositol-free SD medium at 30°C for 20 min. 10 Ci of [ 3 H]inositol were added and the cells incubated for 40 min. 0.75 ml of fresh medium was added, and the incubation continued for 80 min. The yeast was also labeled with C 6 -ceramide complexed with bovine serum albumin as previously described (29). Cells were harvested by centrifugation and washed twice with 1 M sorbitol. Chloroform:methanol (0.4 ml; 1:1 v/v) were added, and cells were disintegrated with glass beads. The pellet was re-extracted several times with chloroform:methanol:water (10:10:3, per volume), the combined supernatants dried, and the lipids prepared and analyzed as previously described (27,28). Again, NBD-ceramide-labeled samples were imaged using the FLA3000 laser fluorescence imaging system (Fuji).

RESULTS
Functional Cloning of the Leishmania AUR1 Orthologue-Candidate genes encoding the putative protozoan protein were selected using a bioinformatic approach based on that used in the identification of the metazoan enzyme SM synthase (29). Briefly, a conserved motif shared by the lipid phosphate phosphatase (LPP) family, fungal AUR1p proteins, and animal sphingolipid synthases (H(YFWH)X 2 D(VLI)X 2 (GA)X 3 (GSTA)) was used to screen the genome sequence data bases of the kinetoplastids. Of the eight candidates identified, one was of unknown function, had sequence orthologues in both Trypanosoma and Leishmania sp., had transmembrane domains consistent with a Golgi-localized IPC synthase, and had no orthologue in mammalian cells (which do not possess an IPC activity). The gene encoding this putative protozoan IPC synthase is present as a single copy in L. major. In contrast, the trypanosomes possess multiple nonidentical loci, two in T. cruzi and four in T. brucei. The Leishmania candidate was amplified and cloned into a URA3-selectable yeast expression vector to create pRS426 LmIPCS. In the auxotrophic yeast strain YPH499-HIS-GAL-AUR1 (see "Experimental Procedures"), the essential AUR1 gene, believed to encode at least part of an IPC synthase (31), is under the control of the GAL1 promoter and therefore is repressed in the presence of glucose. In non-permissive glucose-containing SD medium, transformation with pRS426 LmIPCS restored the growth of YPH499-HIS-GAL-AUR1, as did the ectopic expression of S. cerevisiae AUR1p (ScAUR1) (Fig. 2A). The wild type (YPH499) cannot grow without uracil.
Biochemical analysis of the YPH499-HIS-GAL-AUR1 strain showed that suppression of AUR1 expression by incubation in non-permissive SD medium leads to significantly reduced incorporation of tritiated inositol into base-resistant lipid species. This strongly suggests that IPC synthase activity is downregulated under these conditions (Fig. 2B). The auxotrophic mutant strain transformed with empty pRS426MET25 vector was grown for 20 h in either SD or permissive SGR medium. At this point, both sets of cells were in the stationary phase of growth and 100% viable (as shown by a trypan blue exclusion assay and plating on permissive medium (data not shown)). Furthermore, when grown in non-permissive medium, YPH499-HIS-GAL-AUR1 demonstrated cell clumping (data not shown), perhaps mirroring the pleomorphic defects in cell morphology and division previously observed in AUR1 mutants of S. cerevisiae and Schizosaccharomyces pombe (31,32). Analysis of the complemented YPH499-HIS-GAL-AUR1 line confirmed LmIPCS as a functional orthologue of AUR1p. YPH499-HIS-GAL-AUR1 rescued using either pRS426 ScAUR1 or pRS426 LmIPCS and labeled with [ 3 H]inositol demonstrated a base-resistant sphingolipid profile identical to that observed in the wild type (Fig. 2B). However, in the LmIPCS-complemented line, the relative quantity of labeled mannosylated IPC was significantly diminished. In addition, an accumulation of FIGURE 2. A, rescue of YPH499-HIS-GAL-AUR1 by complementation with either the yeast AUR1 (ScAUR1) or a predicted Leishmania orthologue identified using bioinformatics (LmIPCS). YPH499, wild type; pRS426, YPH499-HIS-GAL-AUR1 pRS426MET25; ScAUR1, YPH499-HIS-GAL-AUR1 pRS426 ScAUR1; LmIPCS, YPH499-HIS-GAL-AUR1 pRS426 LmIPCS. SD, synthetic minimal medium supplemented (or not) with histidine (HIS) and uracil (URA) for selection of the pRS426MET25 plasmid. B, YPH499-HIS-GAL-AUR1 is deficient in inositol sphingolipid biosynthesis when incubated in non-permissive, glucose-containing SD medium; however, complementation with either yeast AUR1 (ScAUR1) or the protozoan orthologue (LmIPCS) open reading frames facilitates synthesis of a base-resistant inositol sphingolipid profile identical to that of wild type yeast. Shown is material extracted from wild type YPH499, IPT1 mutant RCD113 (45), and YPH499-HIS-GAL-AUR1 transformed with pRS426MET25, pRS426 ScAUR1, or pRS426 LmIPCS maintained in the medium indicated and inositol-labeled as described under "Experimental Procedures." The base-resistant lipids (BASE ϩ, left), which co-migrated with base-sensitive phosphatidylinositol species (BASE Ϫ, right) are designated as IPC and are likely to represent species with varying levels of hydroxylation of the sphinganine and fatty acid moieties (7,46). Unlike the other samples, YPH499-HIS-GAL-AUR1 pRS426 was labeled in both permissive (SGR) and non-permissive (SD) medium at the stationary phase of growth. It is likely that this accounts for the minor differences in the labeled profile. The identities of the base-resistant sphingolipid species (left) were established by comparison with the S. cerevisiae IPT1 mutant strain RCD113, which is deficient in mannose (IP) 2 ceramide (M(IP) 2 C) synthesis and accumulates mannose IPC (MIPC) (45). PI, phosphatidylinositol; X, unknown; O, origin; SD, synthetic minimal medium with glucose; SGR, synthetic minimal medium with galactose.
base-sensitive lyso-phosphatidylinositol and lipid X was evident (Fig. 2B). A yeast mutant deficient in ceramide synthesis with a consequently low level of IPC synthase activity exhibits a similar phenotype (33), suggesting that LmIPCS does not function optimally in this system. This conclusion is supported by the relatively slow growth of the LmIPCScomplemented yeast (data not shown). Despite this phenotype, these complementation data establish LmIPCS as a functional orthologue of S. cerevisiae AUR1p, the putative IPC synthase.
The Leishmania AUR1 Orthologue Defines a New Class of Sphingolipid Synthase-Based on primary sequence identity, orthologues of LmIPCS were identified in the related kinetoplastids T. cruzi and T. brucei (see "Experimental Procedures"), suggesting that IPC synthase is a ubiquitous feature of this important group of flagellated parasites. Similar to Leishmania sp., T. cruzi synthesizes IPC as its predominant complex sphingolipid, and this parasite is known to possess IPC synthase activity (17). Therefore, the two closely related T. cruzi-predicted proteins, sharing 52-53% sequence identity and 69 -70% similarity with the Leishmania orthologue, were provisionally designated TcIPCS1 and TcIPCS2. There are no published data indicating whether T. brucei produces IPC or SM. Given this uncertainty, the four predicted orthologues were provisionally designated as T. brucei sphingolipid synthases (TbSLSs) 1-4. The four TbSLS predicted proteins share 43-44% identity and 61-63% similarity with LmIPCS. The high level of conservation within these kinetoplastid species is illustrated in Fig. 3A, together with the seven transmembrane domains identified during topology prediction of LmIPCS and those regions largely conserved with respect to the animal SM synthases (29). SM synthase motifs D3 (CGDX 3 SGHT) and D4 (HYTXDVX 3 YX 6 FX 2 YH) are similar to the C2 and C3 motifs in LPP (29), as are domains 3 and 4 of the fungal AUR1p (34). These regions contain the histidine and aspartate residues (Fig. 3B) that comprise the catalytic triad that mediates nucleophilic attack on lipid phosphate ester FIGURE 3. A, the predicted kinetoplastid AUR1p orthologues are highly conserved. Shown is ClustalW alignment. TM, predicted LmIPCS transmembrane domains; D, conserved SM synthase domains. Shaded black, 100% identical; shaded gray, 100% similarity. B, the predicted kinetoplastid proteins share a high level of similarity with the SM synthases with respect to three domains. Edited ClustalW alignments of the conserved SM synthase domains D1, D3, and D4 equivalent to LmIPCS amino acids 67-70, 212-221, and 264 -284 respectively. Shaded black, 50% identical; shaded gray, 50% similarity. *, conserved histidine and aspartate residues that form the catalytic triad. C, the kinetoplastid proteins define a new class of sphingolipid synthases. An evolutionary tree constructed using maximum parsimony of sequence equivalent to LmIPCS amino acids 77-313, a region conserved between the Metazoa, Fungi, Apicomplexa, and Kinetoplastidae and including the domains within which the amino acids forming the catalytic triad are present. The sequence of the human lipid phosphate phosphatase 1 (LPP1) was the designated out-group. Bootstrap  bonds (29,35). Mutagenesis of this triad has been shown to inactivate fungal IPC synthase activity (36). Huitema et al. (29) predict that the D1 (PLPD) and D2 (RRX 8 YX 2 RX 6 T) motifs are entirely unique to the SM synthases.
Analyzing the conserved domains D1, D3, and D4 in isolation (Fig. 3B), it is clear that the predicted kinetoplastid proteins possess the histidine and aspartate residues forming the cata-lytic triad. However, it is also evident that they demonstrate a higher degree of similarity in these regions with the animal SM synthases than with the fungal AUR1p proteins. This is clearly illustrated in the 100% identity of D1, positioned between the first two transmembrane domains of both the animal and kinetoplastid proteins, a domain with no equivalent in the fungal enzyme. Notably, both mammal (29) and Leishmania (37) sphingolipid synthases utilize ceramide (rather than phytoceramide as in fungi); therefore, the conserved D1 domain may function in the preferential binding of this substrate. Unlike the SM synthases (29), the protozoan D1 is predicted to face the cytosol (Fig. 4D). However, there is no experimental evidence to support this prediction, and furthermore, ceramide is delivered to both the lumenal and cytosolic faces of the Golgi apparatus (38). Conversely, the protozoan proteins do not possess a motif with similarity to SM synthase D2.
These data indicate that the protozoan enzymes are not closely related to fungal AUR1p proteins, despite exhibiting an equivalent function, at least in the case of Leishmania. Further, phylogenetic analysis of both predicted SM and IPC synthase sequences using maximum parsimony demonstrated that the kinetoplastid proteins form a distinct clade and, as such, are proposed to define a new class of sphingolipid synthases (Fig. 3C). This analysis failed to place the putative SM synthases of the malaria parasite Plasmodium falciparum (PfSMS) (29), raising a question over their classification and evolutionary origin.
Functional Analyses of LmIPCS Activity in a Mammalian Cell Line-As described above, mammalian cells do not possess either IPC synthase or IPC (Fig. 1). To assay the activity of LmIPCS in a mammalian system, an epitope-tagged gene copy was ectopically expressed in HEK293 cells and the cells subsequently labeled with [ 3 H]inositol to screen for the synthesis of inositol-containing lipids or NBD-ceramide for the analysis of sphingolipid synthesis. Fractionation of inositol-labeled lipids by thin-layer chromatography demonstrated the presence of a base-resistant inositol-containing lipid co-migrating with Leish-  SEPTEMBER 22, 2006 • VOLUME 281 • NUMBER 38 mania IPC (Fig. 4A). The negative control, HEK293 cells transfected with lacZ, do not synthesize this lipid species. Leishmania sp. IPC predominantly contains C16 and C18 sphinganine and fatty acid moieties (37,39,40), more similar to the sphingomyelin in mammalian cells (41) than the very long-chain fatty acid group predominant in S. cerevisiae IPC species (42). Additional analysis of NBD-ceramide-labeled cells demonstrated that expression of LmIPCS in the mammalian cell line leads to production of a ceramide-containing species that co-migrates with both Leishmania and yeast IPC (Fig. 4B). These results strongly indicate that the Leishmania LmIPCS gene encodes a protein that can constitute IPC synthase activity in a null background.

Protozoan IPC Synthase
AUR1p has been localized to the Golgi apparatus in S. cerevisiae (36), whereas in mammalian cells the two distinct isoforms of SM synthase have been localized to the Golgi and the plasma membrane (29). Immunolocalization of epitope-tagged LmIPCS in HEK293 cells demonstrated that the protozoan enzyme is concentrated within the Golgi apparatus, as identified by the co-localization of a GFP-tagged marker, ARF 1 (Fig.  4C). In addition, minor quantities of the protein also localize to the plasma membrane when the expression levels are high (Fig.  4C). Topology prediction using the PHD package indicates that the residues predicted to form the LmIPCS catalytic triad are orientated toward the Golgi lumen (or the cell surface) in a conformation identical to that predicted for both fungal AUR1p and animal SM synthases (29,36) (Fig. 4D). These observations suggest that, as in animals and fungi, kinetoplastid sphingolipid synthesis takes place in the lumen of the Golgi.
Specific Inhibition of Leishmania IPC Synthase-The expression of functional protozoan orthologues in the mutant yeast system described (Fig. 2) provided an ex vivo system in which to screen the existing fungal IPC synthase inhibitors against the functional Leishmania orthologue. In this study, the diffusion assay system previously described, with the sphingolipid bypass mutant (AGD) as a negative control, was utilized (11,30). The conditional AUR1 mutant strain rescued with ScAUR1 was, similar to the wild type control, sensitive to cycloheximide (an inhibitor of protein translation), myriocin (an inhibitor of serine palmitoyltransferase that mediates the first step in sphingolipid biosynthesis), and aureobasidin A (an inhibitor of fungal IPC synthase (11), to which mutations in AUR1 confer resistance (10)). The mutant rescued with LmIPCS, although sensitive to cycloheximide and myriocin, was resistant to aureobasidin A at 25 M (Fig. 5A). However, by increasing the concentration of the IPC synthase inhibitor 4-fold (to 100 M), sensitivity of LmIPCS to the drug was detected (Fig. 5B). The sphingolipid bypass strain AGD is able to grow without synthesizing sphingolipids and therefore is able to tolerate loss-of-function mutations in both serine palmitoyltransferase and IPC synthase (11). Therefore, AGD is resistant to antifungals targeting either of these enzymes, provided these agents have a specific mode of action (30). The lack of an inhibitory effect of aureobasidin A at 100 M on the growth of AGD, together with its action against the rescued mutant, demonstrates that the drug specifically inhibits LmIPCS as well as ScAUR1p. The LmIPCS-rescued mutant showed increased sensitivity to myriocin (an inhibitor of serine palmitoyltransferase) when compared with either the wild type or the mutant rescued with ScAUR1 (Fig. 5A). The reason for this is unclear, but this observation correlates with the generally fragile nature of the cells complemented with the protozoan orthologue that results in a reduced rate of growth (data not shown).
Aureobasidin A (at 20 M) also exhibited an inhibitory effect against insect stage L. major in cell culture. However, in this system, the inhibitor was nonspecific, as demonstrated by an equivalent activity against parasites without an active sphingolipid biosynthetic pathway (14), in which LmIPCS would be predicted to be redundant (data not shown).

DISCUSSION
Sphingolipids are essential membrane components of eukaryotic cells. In mammalian systems, the major complex sphingolipid is sphingomyelin, whereas in the kinetoplastids, including Leishmania sp., it is inositol phosphorylceramide (IPC) (7). A similar situation exists in fungi where IPC is synthesized utilizing the enzyme IPC synthase. The fungal IPC synthases are believed to be encoded by the AUR1 gene, which has been demonstrated to be essential for viability (32,42). No close relatives of this gene can be identified in the kinetoplastid genome data bases. However, in the study reported here, we have identified the single copy gene encoding the functional AUR1 orthologue from L. major (LmIPCS) utilizing a combination of bioinformatic and functional genetic methods. The protein encoded by this gene shares only limited amino acid sequence identity with fungal AUR1p but has closely related sequence orthologues in both the African and American trypanosomes. Notably, the predicted kinetoplastid AUR1p orthologues demonstrate significant homology to three of the four conserved domains of the recently identified animal SM synthases (29). This indicates that, despite possessing an equivalent function, the protozoan enzymes are not closely related to the fungal IPC synthases (AUR1p) but share more similarity, at least at a primary sequence level, with the animal SM synthases. Indeed, phylogenetic analyses indicated that the kinetoplastid proteins define a new class of sphingolipid synthases clustering with neither the fungal or animal sequences (Fig. 3C). The significance of this observation in evolutionary terms is unclear, although the identification of a plant IPC synthase will facilitate more informative analyses.
In addition to complementing an auxotrophic AUR1 yeast mutant, expression of LmIPCS in a mammalian cell line (null for IPC and IPC synthase activity) led to the synthesis of inositol-and ceramide-labeled lipids that co-migrated with Leishmania and yeast IPC. This strongly suggests that LmIPCS is able to constitute IPC synthase activity in an axenic system. LmIPCS is largely confined to the Golgi apparatus, as indicated by subcellular localization, a situation similar to that for the proposed sphingolipid synthases of both fungi and animals (29,36). The association of LmIPCS with the Golgi membrane is predicted to be mediated by seven transmembrane helices, with the putative active site including the catalytic triad of histidine and aspartate residues orientated toward the lumen as in fungal AUR1p and mammalian sphingolipid synthase (29,36). Thus, it is likely that the protozoan, mammalian, and fungal enzymes possess a conserved mechanism of action similar to that described for LPPs (35), whereby, in the case of the yeast and kinetoplastids, IPC synthase catalyzes the transfer of an inositol phosphate group from phosphatidylinositol to the 1-hydroxyl group of ceramide or phytoceramide, releasing diacyl glycerol as a by-product. This is predicted to occur via a two-step process involving initial transfer of the inositol phosphate residue to an activated histidine in the active site, with the phosphate intermediate being subsequently subjected to nucleophilic attack by the oxygen of the (phyto)ceramide hydroxyl group, resulting in transfer to the sphingoid base. Notably, mutation of His-294, part of catalytic triad in the yeast AUR1p, results in nonviable haploid cells (36). Taken together with the phylogenetic analysis in this study, the predicted conservation of the catalytic mechanism suggests that fungal, animal, and protozoan sphingolipid synthases have evolved into three distinct classes from a common ancestor. Conversely, the soluble bacterial SM synthase lacks the motifs believed to be integral for activity of the eukaryotic enzymes (43), indicating that it arose independently.
The data presented here demonstrate that LmIPCS (like its fungal equivalent) is inhibited by aureobasidin A. Given that IPC synthase activity is detectable in the pathogenic stages of all of the parasitic kinetoplastids studied to date (15,17), the protozoan enzyme may represent a tractable drug target. The toxicity and expense of available treatments for leishmaniasis, African sleeping sickness, and Chagas disease remain major international health problems. These facts, coupled with the prevalence of drug-resistant parasites, make the discovery of novel drug targets a priority. The identification and characterization of the kinetoplastid AUR1p orthologue, a probable IPC synthase with no functional equivalent in mammalian cells, will raise the possibility of developing specific inhibitors with little or no mammalian toxicity. In addition, the close evolutionary relationship observed between the kinetoplastid AUR1p orthologues indicates that the development of broad spectrum inhibitors is a feasible objective, work that may lead to a new generation of anti-protozoals directed against the causative agents of a range of emerging diseases.